PROJECT SUMMARY The first combination immunotherapy of ipilimumab (anti-CTLA-4) and nivolumab (anti-PD-1) approved for melanoma in 2015 has achieved longer progression-free survival compared to nivolumab alone (11.5 vs. 6.9 months), with 53% of patients still alive at 4 years, of whom 71% remain treatment free. Yet, safety is the biggest barrier to broader implementation of this successful combination, due to much greater occurrence of serious immune-related adverse events (irAEs) compared to nivolumab alone. In particular, immune-related colitis (irColitis) is the most common occurrence, with 10-15% of patients developing it upon receiving anti- CTLA-4 alone or in combinations. Such irColitis is life-threatening and requires immediate treatment with high- dose steroids or biologics such as anti-TNF, as well as ipilimumab or immunotherapy discontinuation. These safety concerns have hampered the development of CTLA-4 targeting agents and may have reduced their efficacy due to lower recommended dosing in combination therapy and limited treatment course. There is therefore an urgent need to comprehensively study the nature of irColitis and find whether potential risk factors could be identified and mitigated for greater safety and efficacy of combination treatments. We have assembled a unique team of inflammatory bowel disease (IBD) scientists and clinicians, tumor immunologists, microbiome specialists, and medical oncologists specializing in immunotherapy of melanoma to analyze colorectal tissue lesions, the gut microbiome, and peripheral blood biomarkers that could predict or contribute to development of irColitis. By comparing what we already learned from studying biospecimens of IBD patients, i.e., their genetic subsets at the single cell level, gut microbiome, and predictive immune signatures, we aim to characterize the transcriptomic, immunopathologic, serologic, and gut microbiome landscape of irColitis and leverage this knowledge for better therapeutic options. We will address irColitis by characterizing gut biopsies of melanoma patients during treatment using single- cell RNA sequencing and multiplex immunohistochemistry mapping, and we will relate immune subsets and markers discovered to those found in IBD lesions. We will also measure peripheral blood serologic markers (anti-microbial and anti-GM-CSF autoantibodies by ELISA, soluble cytokine and protein analytes with Olink proximity extension assay) that we recently found to be risk factors detectable prior to development of IBD, and analyze them for their capacity to predict irColitis. From stool collected before and throughout treatment, we will analyze by 16S and metagenomic sequencing the gut microbiome of melanoma patients with or without irColitis to compare it to known colitogenic structure in IBD, and functionally assess bacterial communities in gnotobiotic models. Finally, we will integrate our findings to establish a mechanistic model of irColitis compared to IBD, to propose novel stratifications of cancer immunotherapy patients by risk factors, and to offer future therapeutic opportunities, such as actionable tissue genomic and protein targets or fecal transplants.